Networks, Communities, and the Science of Sciences

This lecture explores the fundamental challenges in network community detection and the science of science. Dive into the limitations of popular clustering algorithms based on global quality function optimization like modularity maximization. Examine the critical issue of validation in network community detection, including the use of realistic benchmark graphs with built-in community structures and the role of metadata. Discover how neural embeddings can efficiently detect communities within networks. The presentation also investigates science as a system through data analysis, revealing how citation distributions across disciplines can be normalized to reveal universal patterns, and examines the COVID pandemic's impact on scientific research. Presented by Santo Fortunato from Indiana University Bloomington, this talk provides valuable insights for those interested in network science, scientometrics, and complex systems analysis.